Anthropic's Claude Mythos AI raises cyber‑insurance concerns

Anthropic introduced Claude Mythos Preview, an AI model capable of autonomously finding and exploiting zero‑day vulnerabilities, prompting industry concern about large‑scale cyber attacks.

Investment manager Man Group says existing cyber catastrophe bonds, mostly per‑occurrence structures with high attachment points, currently shield insurers but may be stressed if AI tools are weaponized. Man Group highlights legal ambiguities in defining a single cyber event and notes potential systemic supply‑chain risk as AI becomes integral to corporate processes.

Anthropic also launched Project Glasswing, bringing together major software vendors to use Mythos Preview for defensive security. Meanwhile, Anthropic’s Claude Opus 4.8 model has suffered recurring outages and elevated error rates since June 2026, disrupting enterprise users and drawing criticism for limited communication from the company.